That Times article

[Tom Crispin]

www.johnballcycling.org.uk/misc/********

What gets me is the graph on the right. It makes cycling look
comparably unsafe to motorcycling - when it is, in fact, 11 times
safer. Also the bar for fatalities per 100,000,000 Km, which shows
cycling to be safer than walking, is less prominent than the other
bars.

I believe that the article is deliberately misleading, and that a
formal complaint would be justified.

 

[Finlay Mackay]

On 2007-10-19, Tom Crispin <[email protected]> wrote:
> www.johnballcycling.org.uk/misc/********
>
> What gets me is the graph on the right. It makes cycling look
> comparably unsafe to motorcycling - when it is, in fact, 11 times
> safer. Also the bar for fatalities per 100,000,000 Km, which shows
> cycling to be safer than walking, is less prominent than the other
> bars.
>
> I believe that the article is deliberately misleading, and that a
> formal complaint would be justified.

Agreed, although I think you have to put the blame on the subeditors
rather than the writer. Most of the article is about motorcyclists, but
this is not reflected in the headline or the graphic. The text is
actually fairly balanced, if you ignore the idiotic premise (casualties
per km is the only sensible measure to use if you want to compare the
relative merits of different modes of transport for the same journey).

cheers

Finlay

 

[John Kane]

On Oct 19, 1:30 pm, Tom Crispin
<[email protected]> wrote:
> www.johnballcycling.org.uk/misc/********
>
> What gets me is the graph on the right. It makes cycling look
> comparably unsafe to motorcycling - when it is, in fact, 11 times
> safer. Also the bar for fatalities per 100,000,000 Km, which shows
> cycling to be safer than walking, is less prominent than the other
> bars.
>
> I believe that the article is deliberately misleading, and that a
> formal complaint would be justified.

It took me a moment to see the axes breaks. The graph is at best the
result of outstanding incompetence.

Using a break like that has been known to be deceptive since Darryl
Huff''s "Lying with Statistics". I wonder if the Times was going for
an entry in a new editon?

A slightly more honest version can be seen at
http://ca.geocities.com/jrkrideau/cycling/times.stats.pdf however
scale makes some of the results look actually less than zero.

 

[Paul Boyd]

Tom Crispin said the following on 19/10/2007 18:30:
> www.johnballcycling.org.uk/misc/********
>
> What gets me is the graph on the right. It makes cycling look
> comparably unsafe to motorcycling - when it is, in fact, 11 times
> safer. Also the bar for fatalities per 100,000,000 Km, which shows
> cycling to be safer than walking, is less prominent than the other
> bars.
>
> I believe that the article is deliberately misleading, and that a
> formal complaint would be justified.

"But The Times requested figures per journey and per hour. These
unpublished numbers....."

It obviously didn't dawn on them that the reason these figures were
unpublished is because they're meaningless!

--
Paul Boyd
http://www.paul-boyd.co.uk/

 

[Nick]

Finlay Mackay wrote:
> On 2007-10-19, Tom Crispin <[email protected]> wrote:
>> www.johnballcycling.org.uk/misc/********
>>
>> What gets me is the graph on the right. It makes cycling look
>> comparably unsafe to motorcycling - when it is, in fact, 11 times
>> safer. Also the bar for fatalities per 100,000,000 Km, which shows
>> cycling to be safer than walking, is less prominent than the other
>> bars.
>>
>> I believe that the article is deliberately misleading, and that a
>> formal complaint would be justified.
>
> Agreed, although I think you have to put the blame on the subeditors
> rather than the writer. Most of the article is about motorcyclists, but
> this is not reflected in the headline or the graphic. The text is
> actually fairly balanced, if you ignore the idiotic premise (casualties
> per km is the only sensible measure to use if you want to compare the
> relative merits of different modes of transport for the same journey).
>

Comparison of risk should give an indication of the total risk to the
person. Riding a bike encourages longer journeys than walking. I ride
150km a week there is no way I would ever walk that far. The choice is
not always between two forms of transport for the same journey.

The times article is good because it gives people the ability to choose
the statistic they apply to their own circumstances.

 

[p.k.]

Finlay Mackay wrote:
> The text is
> actually fairly balanced, if you ignore the idiotic premise
> (casualties per km is the only sensible measure to use if you want to
> compare the relative merits of different modes of transport for the
> same journey).


But that is exactly what was done here a while ago in saying that walking to
school was more dangerous than cycling in the discussion re guidance note to
patents.

pk

 

[Ian Smith]

On Sat, 20 Oct 2007 19:18:48 +0100, p.k. <[email protected]> wrote:
> Finlay Mackay wrote:

> > The text is actually fairly balanced, if you ignore the idiotic
> > premise (casualties per km is the only sensible measure to use if
> > you want to compare the relative merits of different modes of
> > transport for the same journey).
>
> But that is exactly what was done here a while ago in saying that
> walking to school was more dangerous than cycling in the discussion
> re guidance note to patents.

No, the bracketed bit is not teh idiotic premise. At least, that's
how I read it, and it's noty teh premise of teh article (again, as I
read it).

The idiotic premise was that risk-per-journey is a useful measure
when deciding how to undertake a particular journey.

regards, Ian SMith
--
|\ /| no .sig
|o o|
|/ \|

 

[Ian Jackson]

In article <[email protected]>,
John Kane <[email protected]> wrote:
>A slightly more honest version can be seen at
>http://ca.geocities.com/jrkrideau/cycling/times.stats.pdf however
>scale makes some of the results look actually less than zero.

Another problem with the Times's graph which you haven't corrected is
that the three different ways of comparing - per journey, per km and
per hour - are presented with arbitrary relative scaling on the same
graph. The magnitude axis isn't in the same units for each comparison
method, although expressing it as `per 100,000' obscures this.

Furthermore, the choice of specific units and scales causes the `per
hours' and `per journey' graphs to use more of the available space.
The effect is as if there were `more' deaths `per hour' than `per km',
which is obviously nonsense - but it leads the eye to think that these
numbers are the really serious and important ones.

A fairer representation would be something like this:

per km per journey per hour
Bus/coach | | |
Car |= |= |
Foot |#####= |= |
Pedal Cycle |##### |#= |
Motorcycle |############### |############### |###############
0 5 0 1 0 3
deaths/ 100M km /1M journeys /1M hours

The inclusion of motorcycles which are far more dangerous by all
measures completely obscures the coomparisons. Without motorcycles
you get this:

per km per journey per hour
Bus/coach | |= |
Car |# |#### |####
Foot |############### |#### |######
Pedal Cycle |############## |############### |###############
0 1 2 0 1 2 3 0 1 2 3
deaths/ 100M km /10M journeys /10M hours

Of course that still doesn't fairly represent the effect of cycling on
life expectancy because it neglects the health benefits and of course
`per journey' is a ridiculous approach anyway.

NB that you will obviously need a fixed-width font to make sense of
the above.

--
Ian Jackson personal email: <[email protected]>
These opinions are my own. http://www.chiark.greenend.org.uk/~ijackson/
PGP2 key 1024R/0x23f5addb, fingerprint 5906F687 BD03ACAD 0D8E602E FCF37657